Marine Ecology Progress Series 491:165

Home , Pyrosome, Salp, Thetys vagina, Total carbon

Vol. 491: 165–175, 2013 MARINE ECOLOGY PROGRESS SERIES Published October 2 doi: 10.3354/meps10450 Mar Ecol Prog Ser

Salp-falls in the Tasman Sea: a major food input to deep-sea benthos

Natasha Henschke1,2,*, David A. Bowden3, Jason D. Everett1,2,4, Sebastian P. Holmes5,6, Rudy J. Kloser7, Raymond W. Lee8, Iain M. Suthers1,2

1Evolution and Ecology Research Centre, University of New South Wales, Sydney, New South Wales 2052, Australia 2Sydney Institute of Marine Science, Building 22, Chowder Bay Road, Mosman, New South Wales 2088, Australia 3National Institute of Water and Atmospheric Research Ltd. (NIWA), 301 Evans Bay Parade, Greta Point, Wellington 6021, New Zealand 4Plant Functional Biology and Climate Change Cluster, Faculty of Science, University of Technology Sydney, PO Box 123 Broadway, New South Wales 2007, Australia 5School of Biological Sciences, University of Sydney, Sydney, New South Wales 2006, Australia 6Water & Wildlife Ecology Group (WWEG), School of Science & Health, University of Western Sydney (UWS), Penrith, New South Wales 1797, Australia 7Commonwealth Scientific and Industrial Research Organisation (CSIRO) Marine Laboratories, PO Box 1538, Hobart, Tasmania 7001, Australia 8School of Biological Sciences, Washington State University, PO Box 644236, Pullman, Washington 99164-4236, USA

ABSTRACT: Large, fast-sinking carcasses (food-falls) are an important source of nutrition to deep- sea benthic communities. In 2007 and 2009, mass depositions of the salp Thetys vagina were observed on the Tasman Sea floor between 200 and 2500 m depth, where benthic crustaceans were observed feeding on them. Analysis of a long-term (1981 to 2011) trawl survey database determined that salp biomass (wet weight, WW) in the eastern Tasman Sea regularly exceeds 100 t km−3 yr−1, with biomasses as high as 734 t km−3 recorded in a single trawl. With fast sinking rates, salp fluxes to the seafloor occur year-round. Salps, like jellyfish, have been considered to be of low nutritional value; however, biochemical analyses revealed that T. vagina has a carbon (31% dry weight, DW) and energy (11.00 kJ g−1 DW) content more similar to that of phytoplankton blooms, copepods and fish than to that of jellyfish, with which they are often grouped. The deposition of the mean yearly biomass (4.81 t km−2 WW) of salps recorded from the trawl database in the Tasman Sea represents a 330% increase to the carbon input normally estimated for this region. Given their abundance, rapid export to the seabed and high nutritional value, salp carcasses are likely to be a significant input of carbon to benthic food webs, which, until now, has been largely overlooked.

KEY WORDS: Benthic communities · Gelatinous zooplankton · Carbon cycling · Fluxes · Salp-fall · Jelly-fall

Resale or republication not permitted without written consent of the publisher

INTRODUCTION including phytoplankton blooms, other plant or algal matter, zooplankton faecal pellets and carcasses of Food-limited deep-sea benthic ecosystems rely on larger fauna are major contributors of organic matter depositions of organic matter from the euphotic zone to the sea floor (Rowe & Staresinic 1979, Smith et al. (Gooday 2002). Concentrated pulses of particulate 2008). Despite the majority of particles being small organic matter (POM) derived from differing sources (<5 mm) (Alldredge & Silver 1988), these pulses are

*Email: [email protected] © Inter-Research 2013 · www.int-res.com 166 Mar Ecol Prog Ser 491: 165–175, 2013

an important source of nutrition for deep-sea benthic al. 1989, Lyle & Smith 1997, Gili et al. 2006), they are communities, promoting both species richness and still generally thought to be of low nutritional value abundance (Butman et al. 1995). Benthic ecosystem (Moline et al. 2004). Therefore, to determine whether functions are also positively related to increasing salp carcasses can positively contribute to the ben- POM supply, including sediment community respira- thic ecosystem, it is necessary to identify the quality tion rates and organic matter remineralisation (Witte of food they provide. & Pfannkuche 2000, Smith et al. 2008, Sweetman & In particular, we sought to (1) assess the frequency Witte 2008). and abundance of salp swarms in the Tasman Sea Large, fast-sinking particles, such as carcasses, and eastern New Zealand over 30 yr, (2) quantify the provide food-fall events that augment the nutritional biomass and relative abundance of Thetys vagina ecology of deep-sea benthic communities (Rowe & carcasses on the sea floor, and (3) compare the ener- Staresinic 1979, Stockton & DeLaca 1982, Smith & getic input and the biochemical composition of T. Baco 2003). The ‘gelatinous pathway’ (Billett et al. vagina carcasses with other gelatinous zooplankton. 2006, Lebrato et al. 2012) was first discovered by Moseley (1880) and illustrates the potential for sinking carcasses of gelatinous organisms to contribute a MATERIALS AND METHODS large flux of organic matter to the benthic environment. Because of their swarming nature, depositions Study region of gelatinous carcasses generally accumulate in high densities to the benthic environment in areas under- Long-term trawl surveys and 2 benthic sampling lying large and persistent gelatinous populations cruises were conducted in the southern Tasman Sea (Billett et al. 2006, Lebrato & Jones 2009). For ex - and Pacific Ocean east of New Zealand (Fig. 1A). For ample, following swarms in surface waters (Wiebe et the first benthic study on board the RV ‘Tangaroa’ in al. 1979, Grassle & Morse-Porteous 1987), dense con- June 2007 (TAN0707), sampling was carried out on centrations of salp carcasses were observed nearby the Challenger Plateau, a large submarine plateau on the seafloor in the outer Hudson Canyon (3240 m) extending from the west coast of central New Zea - in 1975 and 1986 (Cacchione et al. 1978). Similarly, land and considered to be a region of low pelagic pelagic cnidarian deposits (jelly-falls) have been re- productivity (Wood 1991). In October 2009, on the corded on the sea floor off Oman (Billett et al. 2006), second benthic study on board the RV ‘Southern Sur- in the Sea of Japan (Yamamoto et al. 2008) and in a veyor’ (SS03/2009), sampling occurred off southeast- Norwegian fjord (Sweetman & Chapman 2011), while ern Australia in Bass Canyon, one of the largest sub- pyrosome carcasses have been observed on the marine canyons in the world (Mitchell et al. 2007). Madeira Abyssal Plain (Roe et al. 1990) and on the seafloor off the Ivory Coast (Lebrato & Jones 2009). During 2 benthic sampling research voyages, we Trawl data analysis observed mass depositions of the large salp Thetys vagina on the Tasman Sea floor, prompting an exam- Trawl data were available from 2 sources: a long- ination into their subsequent fate and the nutritional term data series (30 yr) from the New Zealand fish- value provided by the carcasses to the deep-sea ben- eries research trawl database and pelagic trawls in thic communities. T. vagina reaches up to 306 mm in the Tasman Sea over 3 yr. Salp and pyrosome bio- size (Nakamura & Yount 1958) and has a distribution mass was obtained from analysis of the New Zealand spanning the top 200 m (Thompson 1948, Iguchi & fisheries database (stock assessment, research and Kidokoro 2006) of sub-tropical and temperate waters observer-monitored commercial trawls) from 1981 to of the Mediterranean Sea and the Atlantic, Indian 2011 (n = 2044; see Fig 1A for sampling locations). As and Pacific oceans (Berrill 1950). Salp carcasses can the majority of data was opportunistically sampled, potentially sink at rates of up to 1700 m d−1 (Lebrato sampling periods within a year are variable but on et al. 2013), suggesting that little, if any, decomposi- average include every month per year. Trawls (mid- tion occurs during descent, and mass depositions of water or benthic) were towed at a mean depth of salp carcasses may represent an important and sub- 563.17 ± 342.40 m, ranging from 33 to 2532 m. Where stantial food-fall event for the benthic ecosystem. possible, recorded trawl dimensions and tow lengths Although several reports indicate that gelatinous were used. If details of trawl size or tow distance organisms such as salps are important to the diet of were not available, a standard averaged value calcu- some marine organisms (e.g. Duggins 1981, Clark et lated from all trawls was used (headline height = 8 m, Henschke et al.: Salp carcasses as food-fall 167

30° S A Australia Sydney 35°

Tasman Sea Auckland New Zealand 40°

Wellington Hobart

45° 1981−2011 Salp and pyrosome trawls 2008−2011 T. vagina trawls 2007 Challenger Plateau T. vagina deposition 2009 Bass Canyon T. vagina deposition 50° 140°E 145° 150° 155° 160° 165° 170° 175° 180° 00 38° 36° 2 S B S C Salps per 1000m −2000 No count Australia 1 −1000 −200 10 38° −600 100

−1000 −2000 − −1000 38.5° 2000 −600 −600 Salps per 1000m 2 − 40° 1000 50 250

500 New 200 − Zealand 39° 42° 147°E 148° 149° 150° 166°E 168° 170° 172° Fig. 1. (A) Survey area in the southern Tasman Sea and southwestern Pacific Ocean east of New Zealand showing trawl stations and benthic sampling stations. (B,C) Density distribution based on video footage/camera stills of Thetys vagina (ind. 1000 m−2) at different stations (B) in Bass Canyon and (C) on the Challenger Plateau. Depth contours are displayed in metres wing distance = 30 m, tow distance = 4.4 km, tow conservative than data obtained from the trans- speed = 6.5 km h−1). Individuals were not classified Tasman pelagic trawls. All biomass estimates are into species. Thetys vagina biomass was obtained represented in wet weight (WW). from 3 trans-Tasman cruises in 2008, 2009 and 2011 (n = 12). Depth-stratified midwater tows with a pelagic trawl were made at 200 m intervals to a max- Benthic sample collection and analysis imum depth of 1000 m from the surface, with equal 20 min tows at 6.5 km h−1. The biomass estimates of At both benthic sampling locations, video surveys T. vagina from the trans-Tasman pelagic trawls were were conducted using towed camera platforms with calculated from the net area with the smallest mesh video and still image cameras. All individual salps size capable of capturing them (minimum 40 mm observed on the seabed were counted along the full mesh). Graded mesh area information was not avail- length of each video transect. If necessary, still cam- able for the nets used in the New Zealand fisheries era images taken every 2 min along the transects database, and as a result, biomass estimates are more were used to aid in identification of individuals. 168 Mar Ecol Prog Ser 491: 165–175, 2013

Deployments lasted from 30 to 60 min, at speeds of al. (1956) with D-glucose as the standard. Energetic 0.25 to 0.50 ms−1. On the Challenger Plateau, 46 values of the salps were determined with a Parr deployments of the Deep Towed Imaging System 6200 isoperibol calorimeter using a benzoic acid stan - (Hill 2009) were conducted at depths ranging from dard and as per the manufacturer’s instructions (Parr 237 to 1831 m. Both video and still cameras were Instrument Company 2008). oriented directly downwards, to facilitate scaling, Carbon and nitrogen contents were measured by and video frame width was calculated in ImageJ combusting the material and using gas chromatogra-

(http://rsbweb.nih.gov/ij/) by measuring widths of phy to separate the resulting N2 and CO2 gases. The approximately 100 frame grabs using the camera’s gases were then analysed with an Isoprime isotope paired lasers (20 cm apart) as a reference. In Bass ratio mass spectrometer to give total carbon and Canyon, the Benthic Optical and Acoustic Grab Sys- nitrogen content. An average of the carbon content tem (Sherlock et al. 2010) was deployed at 3 depths: (n = 68, 31.35% DW) per salp for both locations was 450, 650 and 1500 m. As the camera system did not used to calculate carbon standing stock (mg C m−2) have paired lasers, video frame width was measured from the carcasses observed. All salps viewed in the by using the average (± SD) length of Thetys vagina video transects were assumed to have a DW of 0.38 g species caught from the subsequent Bass Canyon (the mean of n = 27 weighed individuals), allowing trawls (55.66 ± 5.90 mm, n = 30) to approximate frame carbon standing stock to be calculated per square size from 17 randomly chosen screenshots containing metre. While this is an approximation, we are confi- T. vagina. Abundance of individuals per 1000 m2 dent that all carcasses seen in the video and captured (ind. 1000 m−2) was calculated by determining salps in benthic gear were of similar size. per corrected area of deployments (corrected area = transect seabed area × percentage of usable video footage). Video analyses were run in Ocean Floor Ob - RESULTS servation Protocol (http://ofop.texel.com); see methods in Bowden et al. (2011). Still image analyses used Observations of Thetys vagina on the sea floor ImageJ software. After each towed camera transect, benthic fauna Carcasses of Thetys vagina were observed in all 3 were sampled at the same site using either a beam video transects in Bass Canyon and in 38 out of 46 trawl (4 m mouth width, 10 mm mesh) or an epiben- transects on the Challenger Plateau (Fig. 2A). In thic sled (1 m mouth width, 25 mm mesh). Trawls total, 368 carcasses were recorded in Bass Canyon were towed for approximately 15 min at 0.75 m s−1. comprising 47.8% of the total observed fauna over an Once back on deck, all fauna were sorted into species, area of 2118 m2. The mean (± SD) density of T. vagina weighed for biomass estimates and frozen (−20°C). was 219 ± 168 ind. 1000 m−2, with a minimum density Salps were thawed, and total length and wet weight of 85 and a maximum of 408 ind. 1000 m−2 (Fig. 1B). were measured for each individual. Guts were On the Challenger Plateau, 1400 individuals were removed prior to biochemical analysis to ensure that observed, making up 9.8% of total observed fauna only body tissue was analysed. Randomly selected over an area of 72 995 m2. In 11 transects where individuals from each site were then freeze-dried abundances of T. vagina were high (>20 ind. 1000 m−2, and their dry weights (DW) recorded. To determine Fig. 1B), T. vagina carcasses ranged from 19.6 to ash-free dry weight (AFDW) of the specimens, tissue 48.7% of the total fauna observed, similar to that samples were taken and combusted at 550°C for 24 h. found in Bass Canyon. The mean (± SD) density of T. All remaining tissue was ground in a ball mill to give vagina on the Challenger Plateau was 26 ± 39 ind. a homogenous powder for biochemical analyses. 1000 m−2, significantly lower than densities found in

Bass Canyon (p < 0.001, F1,48 = 40.1, ANOVA), with a minimum density of 0 and a maximum of 202 ind. Biochemical analyses 1000 m−2 (Fig. 1C). T. vagina comprised 19.0% of total haul biomass on the Challenger Plateau and Protein content of the salps was measured using 42.6% of total haul biomass in Bass Canyon and was the Bradford protein assay (Bradford 1976) with the dominant organism in both locations (see Appen- bovine serum albumin as the standard. Lipid content dix 1). During one transect on the Challenger Pla- of the salps was estimated using a chloroform: teau, the deep-water spider crab Platymaia maoria methanol procedure after Folch et al. (1957) and car- was twice observed directly feeding on T. vagina carbohydrate content was estimated following Dubois et casses (Fig. 2B; Table 1). On 17 occasions across 9 Henschke et al.: Salp carcasses as food-fall 169

Table 1. Megafaunal taxa observed directly feeding on or close to (potential feeders) Thetys vagina carcasses on the Challenger Plateau. Footnotes b to e denote previous records of taxa feeding on salps

No. of events

Crustacea Platymaia maoria 2a Fish Coelorinchus sp.b 10 Paraulopus sp. 1 Tripterophycis gilchristi 1 Helicolenus sp.c 1 Hoplichthys haswelli 1 Hydrolagus novaezelandiaed 1 Echinodermata Ophiuroidea 2 Asteroideae 1 aDirect feeding observed; bClark (1985); cBax & Williams (2000); dDunn et al. (2010); eDomanski (1984)

imately half of the years sampled (Fig. 3A). Biomass ranged from 0.006 t km−3 WW (0.003 t km−2) to 1464 t km−3 WW (824 t km−2), with a 30 yr average (±SD) of 8.54 ± 51.79 t km−3 WW (4.81 t km−2). Salps and pyrosomes were present year-round but appear to form dense swarms an order of magnitude greater than their normal occurrence between December and June (Fig. 3B). High densities of Thetys vagina were captured in 3 trans-Tasman cruises in 2008, 2009 and 2011 (Fig. 3A), with a maximum of 734 t km−3 WW (147 t km−2) caught in 2009 (minimum = 0.003 t km−3 WW, mean (±SD) = 44.82 ± 158.20 t km−3 WW). Depth-stratified Fig. 2. Sea floor photographs from the Tasman Sea. Scale sampling showed that 98% of T. vagina biomass bars = 10 cm. (A) Thetys vagina carcasses at 1565 m depth occurred in the top 200 m of the water column. taken in Bass Canyon. (B) Platymaia maora feeding on T. vagina carcass at 482 m on the Challenger Plateau Biochemical composition of Thetys vagina transects, demersal fish and sea stars were recorded near the carcasses (Table 1). The most common dem- Lipids accounted for the highest proportion of ersal fish were rattails Coelorinchus spp. and were macronutrients, making up a mean (±SD) of 10.5 ± found close to the carcasses on 10 occasions. At both 2.8% DW (Table 2). Protein constituted 3.4 ± 1.5% locations, all T. vagina individuals observed on the DW, and carbohydrates constituted 4.4 ± 1.9%. The sea floor were dead, whole and with no visible bacte- mean (± SD) energetic content of Thetys vagina was rial mats or biofilms. 11.0 ± 1.4 kJ g−1 DW. AFDW was high, ranging from 33 to 88% DW, and total organic content of T. vagina represented only 31% of AFDW. Abundance of Thetys vagina and other large salps Mean (± SD) carbon content for Thetys vagina and pyrosomes in the Tasman Sea (31.4 ± 5.4% DW) was much higher than nitrogen content (2.8 ± 1.1% DW; Table 2). Carbon standing Analysis of the New Zealand fisheries database from stock of the T. vagina deposition in Bass Canyon was 1981 to 2011 determined that salp and pyrosome bio- 26.1 mg C m−2. On the Challenger Plateau, carbon mass exceeded 100 t km−3 WW (56 t km−2) in approx- standing stock was lower, with a mean of 3.1 mg C m−2 170 Mar Ecol Prog Ser 491: 165–175, 2013

Table 2. Thetys vagina. Biochemical and elemental composition. n = These mean densities of T. vagina (26 and number of individuals measured 219 ind. 1000 m−2 for the Challenger Plateau and Bass Canyon, respectively) are much n Mean ± SD Range greater than those found for depositions of Protein content (% DW) 68 3.42 ± 1.46 1.10−7.34 the giant jellyfish Nemo pi lema nomurai in Lipid content (% DW) 31 10.50 ± 2.77 6.19−16.48 the sea of Japan (1.1 ind. 1000 m−2) Carbohydrate content (% DW) 18 4.36 ± 1.92 1.34−7.77 (Yamamoto et al. 2008) and the deep-sea Energetic content (kJ g−1 DW) 9 11.00 ± 1.38 8.91−13.33 Carbon content (% DW) 68 31.35 ± 5.34 18.77−42.68 scyphozoan Periphylla periphylla in a Nor- Nitrogen content (% DW) 68 2.82 ± 1.13 1.52−8.09 wegian fjord (10 ind. 1000 m−2) (Sweetman & C:N 68 12.03 ± 3.03 4.73−19.05 Chapman 2011). Densities were similar to those of Pyrosoma atlanticum carcasses off but reaching 24.1 mg C m−2 at some stations. Using the Ivory Coast (70.6 ind. 1000 m−2) described by C:N ratio and energetic content as an indicator of nu- Lebrato & Jones (2009). The mass depositions of fresh tritional quality, T. vagina is nutritionally similar carcasses observed in this study indicate the recent to phytoplankton (Fig. 4). Values for T. vagina are demise of swarms at both locations. On the Chal- much greater than reported values for cnidarians and lenger Plateau, high densities of T. vagina were ctenophores. observed at the surface during sampling, suggesting that the swarm may still have been developing for several weeks after sampling. Sampling during an DISCUSSION ongoing swarm may limit the accuracy of the deposition densities, as some salp and pyrosome species are Observations of Thetys vagina on the sea floor known to migrate to sea floor depths (Roe et al. 1990, Gili et al. 2006). As T. vagina mainly occurs in the top Densities of Thetys vagina carcasses observed on 200 m of the water column and all carcasses viewed the sea floor in this study are among the highest on the video were dead or moribund, it is unlikely recorded for any gelatinous zooplankton deposition. that deposition abundances were overstated.

Salps and pyrosomes Thetys vagina 10000 A 54 98 152 54 12 1000 22 26 63 60 69 208 WW) 52 64 109 -3 9 18 100 1 182 73 61 347 2 9 10 57 50 10 30 4 6 1 1 6 0.1 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011

1000 B 110 184 98 211

WW)100 Biomass (t km 221 112 285 -3 130 113 10 133 196 251

0.1 Biomass (t km Jul Jan Feb Mar Apr May Jun Aug Sep Oct Nov Dec Fig. 3. (A) Maximum yearly biomass for salps and pyrosomes (dark grey bars; New Zealand fisheries database) from 1981 to 2011 and Thetys vagina (white bars; trans-Tasman pelagic trawls) in 2008, 2009 and 2011. Yearly mean (+SD) is represented by the solid line. Number of trawls per year is indicated above each bar. Station locations are presented in Fig. 1. (B) Mean (+SD) monthly biomass of salps and pyrosomes from 1981 to 2011. Number of trawls per month is indicated above each bar. WW = wet weight Henschke et al.: Salp carcasses as food-fall 171

exceed zooplankton biomass by 300%. Similarly, Young et al. (1996) sampled zooplankton in the Tas- man Sea from 1992 to 1994 and found salps on average made up 30% of zooplankton biomass across the 3 yr and at some times up to 90%. To put salp biomass into perspective, hoki Macruronus nova - ezealandiae constitutes New Zealand’s largest fishery (O’Driscoll 2004), with biomass estimated to be 1.2 t km−2 WW (based on an 8 yr average) and representing 97% of all fish biomass in the midwater depth range (Bull et al. 2001). The mean 30 yr average of large salp biomass (4.81 t km−2 WW) for the Tasman Sea and New Zealand region not only exceeds this value but also indicates the prevalence of salp swarms in the Tasman Sea.

Fig. 4. Relationship between mean (±SD) energetic content Biochemical composition of large salps and mean (±SD) C:N ratio as an indicator of quality of different marine organisms as a food item. Values for Thetys vagina obtained from this study. Other values obtained from This study provides the first data on the biochemi- previous studies: phytoplankton (Platt & Irwin 1973), cope- cal composition of Thetys vagina. Results obtained pods (Donnelly et al. 1994, Ikeda et al. 2006), cnidarians are within expected ranges observed for other large and ctenophores (Clarke et al. 1992) and fish (Childress & salps (Madin et al. 1981, Clarke et al. 1992, Dubis- Nygaard 1973). DW = dry weight char et al. 2006). Similar to our salps, higher propor- tions of lipids to protein are found in the Antarctic Abundance of Thetys vagina and other large salps species Salpa thompsoni (5.7 to 6.8% DW) (Dubis- in the Tasman Sea char et al. 2006), while the opposite trend is observed for North Atlantic salp species: 0.96, 0.25 and 0.97% Other large salps such as Salpa thompsoni (Nishi- DW for Pegea confoderata, S. cylindrica and S. max- kawa et al. 1995, Perissinotto & Pakhomov 1998) and ima, respectively (Madin et al. 1981). Differences in S. aspera (Wiebe et al. 1979, Madin et al. 2006) fre- the biochemical composition of salp species are likely quently form large swarms, but records of Thetys to arise from either differing lipid concentrations vagina are sparse. The largest swarm recorded of T. within food sources (Larson & Harbison 1989) or en- vagina occurred in 2004 in the Sea of Japan, with vironmental conditions inciting higher storage of biomasses as high as 900 t km−3 WW (Iguchi & lipids in cooler waters (Dubischar et al. 2006). Previ- Kidokoro 2006), comparable to the maximum of 734 t ous studies show that carbohydrate contents for salps km−3 WW recorded in this study. As New Zealand are generally low (0.8 to 1.3% DW) (Madin et al. fisheries surveys were designed for the capture of 1981, Clarke et al. 1992, Dubischar et al. 2006); how- large pelagic and demersal fish, they are likely to ever, this study recorded levels similar to those of under-represent the true abundances of salps. Trawls protein. These higher values are consistent with ex - would only spend approximately 35% of their time in pected results, as the salp tunic is mainly comprised of the 0 to 200 m depth range that is preferred by the proteins and polysaccharides (Godeaux 1965). majority of large salps in the Tasman Sea (Thompson Although the total organic content (lipids, proteins 1948). Regardless, abundances of salps across the and carbohydrates) of an organism should equal its 30 yr dataset indicate that salp biomass in the Tas- AFDW (Madin et al. 1981), high values of AFDW are man Sea often exceeded 100 t km−3 WW, which is characteristic for gelatinous zooplankton because of considerably higher than previously thought. difficulties in removing ‘water of hydration’ (Madin Tranter (1962) recorded an average zooplankton et al. 1981) when freeze-drying. Similar AFDW val- biomass (excluding salps) of 36 t km−3 WW from 1959 ues have been found for other salps: 27 to 62.7% DW to 1961 in the Tasman Sea, with salps accounting for for Salpa thompsoni (Huntley et al. 1989, Donnelly et an additional 53 t km−3 WW. Maximum swarm values al. 1994) and 66.4% DW for S. fusiformis (Clarke et from the present study show that large salp and pyro- al. 1992), with total organic contents ranging from 19 some swarms in the Tasman Sea can frequently to 51% of AFDW (Madin et al. 1981, Dubischar et al. 172 Mar Ecol Prog Ser 491: 165–175, 2013

2006). Apart from residual water, the most likely including sea stars (Domanski 1984), sea urchins causes for the ‘missing’ compounds are those missed (Duggins 1981), octocorallians (Gili et al. 2006), by the methodology. For example, as the nitrogen mushroom corals (Hoeksema & Waheed 2012) and, content of protein can be assumed to be 16% (protein from this study, the deep-water spider crab Platymaia = N × 6.25) (Madin et al. 1981), from the nitrogen maora have been observed feeding on salps. Simi- values recorded here, protein content should have larly, pyrosome carcasses have provided food for a been as high as 17.6% DW, 4 times higher than our range of megafauna including crustaceans, arthro- detected values. Similar problems detecting proteins pods, anemones and echinoderms (Roe et al. 1990, in gelatinous zooplankton have been seen in previ- Lebrato & Jones 2009), while anemones, shrimp, ous studies (Clarke et al. 1992, Dubischar et al. 2006) crabs and molluscs have been observed near and and are thought to arise from problems with detect- feeding on cnidarian carcasses (Yamamoto et al. 2008, ing cross-linked proteins. Sweetman & Chapman 2011). As salp carcasses can sink at rates up to 1700 m d−1 (Lebrato et al. 2013), they will be able to reach the seafloor in less than 2 to Contribution to the benthic food web 3 d, before significant bacterial degradation can take place. Preliminary experimental data suggest that at Energetic content of Thetys vagina was higher than seafloor temperatures (4°C), Thetys vagina indivi - that of cnidarians and ctenophores (4.35 to 10.17 kJ duals will retain 68% of their mass after 28 d (N. g−1 DW) (Percy & Fife 1981) and other pelagic tuni- Henschke unpubl. data). These results are slower cates such as Pyrosoma atlanticum (4.94 ± 1.55 kJ g−1 than a model-calculated decomposition time of DW) (Davenport & Balazs 1991) and almost as high approximately 20 d for a gelatinous organism, which as some crustacean species (14.77 ± 1.67 kJ g−1 DW) in cludes more labile cnidarians (Lebrato et al. 2011). (Wacasey & Atkinson 1987). Of all gelatinous zoo- As no bacterial mats or biofilms were observed on plankton studied to date, carbon content for T. va- any of the T. vagina individuals viewed or collected gina was second only to P. atlanticum (Davenport & in this study, slow decomposition rates of T. vagina Balazs 1991, Lebrato & Jones 2009). The energetic would allow carcasses to remain on the sea floor until content and C:N ratios suggest that T. vagina car- scavenged or eventually remineralised via the micro- casses have higher food value than other gelatinous bial loop. zooplankton (cnidarians and ctenophores) (Fig. 4) and nutritionally are more similar to the phytoplankton blooms that normally sustain benthic communi- Potential carbon standing stock ties (Rowe & Staresinic 1979, Smith et al. 2008) as well as fish and copepods. As only the tunic of T. Studies in the world’s oceans (Smith & Kaufmann vagina was analysed, nutritional quality has not been 1999), including the Pacific Ocean near New Zealand elevated by gut contents. Compared to smaller salps, (Nodder et al. 2003), have identified that food de- the tunic of T. vagina is relatively thick and com- mand in the benthic community (sediment commu- posed of densely packed fibrous material (Hirose et nity oxygen consumption) often exceeds food supply al. 1999), possibly resulting in elevated nutritional (POM). Salp carcasses are not detected by traditional values. Based on maximum salp biomass values of methods of sampling water column nutrient fluxes, 100 t km−3 WW, these deposition events can poten- such as sediment traps (Lebrato & Jones 2009), and tially export up to 616 GJ km−2 of energy, or 16 t km−2 consequently are not included in current carbon of carbon, to the Tasman Sea benthos every year. budget calculations, resulting in a considerable Several fish species feed exclusively on salps or under estimation of the total flux. Hence, salp car- have salps as a major component of their diets. These casses may be supplementing the smaller POM that species tend to be opportunistic bentho-pelagic feed- can be collected by sediment traps, providing an ers, such as the black oreo Allocyttus niger, smooth extra source of nutrition for the benthic community. oreo Pseudocyttus maculatus, spiky oreo Neocyttus Since particles that generally make up the majority of rhomboidalis, carinate rattail Macrourus carinatus measured carbon flux in the Tasman Sea are <1 mm and small-scaled brown slickhead Alepocephalus (Kawahata & Ohta 2000), these salp deposition australis (Clark et al. 1989, Lyle & Smith 1997). Our events provide a substantial contribution of much results suggest that the salp carasses often found in larger carbon parcels to the benthos. As swarms of these fish guts may result from scavenging at the Thetys vagina were still in surface waters during seafloor. Apart from fish, other benthic feeders sampling on the Challenger Plateau, by the time the Henschke et al.: Salp carcasses as food-fall 173

entire population had collapsed, the input from both eries, Land Information New Zealand, NIWA and the New faecal pellets and carcasses would have been consid- Zealand Department of Conservation. The Australian samples were collected during a Next Wave transit voyage erably higher than values estimated here. (SS03-2009) funded by the Marine National Facility. This Depositions of Thetys vagina on the Challenger is contribution 110 from the Sydney Institute of Marine Plateau in this study only represented 0.19% of the Science. regional annual carbon flux, whereas carbon provided from the Bass Canyon deposition was 10-fold LITERATURE CITED greater, representing 1.5% of the annual flux (Kawa- hata & Ohta 2000). Although gelatinous zooplankton Alldredge AL, Silver MW (1988) Characteristics, dynamics depositions can occur across all bottom topographies, and significance of marine snow. Prog Oceanogr 20: studies have identified much greater biomasses and 41−82 Bax N, Williams A (2000) Habitat and fisheries production in carbon inputs when organisms are in environments the south east fishery ecosystem. Final Report to the that promote concentration, such as canyons or struc- Fisheries Research and Development Corporation, Pro- tures like pipelines (Cacchione et al. 1978, Lebrato & ject No. 94/040, CSIRO Marine Research, Hobart Jones 2009). Carbon standing stocks for cnidarian Berrill NJ (1950) Order Salpida (Demomyaria). In: The Tuni- cata, with an account of the British species, Vol 133. Ray carcasses in the Arabian Sea have been reported as Society, London, p 286−301 −2 high as 78 g C m in some areas, an order of magni- Billett DSM, Bett BJ, Jacobs CL, Rouse IP, Wigham BD tude higher than mean annual flux (Billett et al. (2006) Mass deposition of jellyfish in the deep Arabian 2006), and 22 g C m−2 has been reported for Pyro- Sea. Limnol Oceanogr 51:2077−2083 soma atlanticum carcasses off the Ivory Coast, 13 Bowden DA, Schiaparelli S, Clark MR, Rickard GJ (2011) A lost world? Archaic crinoid-dominated assemblages on times greater than the annual flux (Lebrato & Jones an Antarctic seamount. Deep-Sea Res II 58:119−127 2009). Future studies may benefit from incorporating Bradford MM (1976) A rapid and sensitive method for the bottom topographies when calculating the potential quantitation of microgram quantities of protein utilizing for gelatinous organisms to accumulate on the sea the principle of protein-dye binding. Anal Biochem 72: 248−254 floor and their eventual contribution to carbon fluxes Bull B, Livingston ME, Hurst R, Bagley N (2001) Upper-slope in the area. fish communities on the Chatham Rise, New Zealand, 1992–99. NZ J Mar Freshw Res 35:795−815 Butman CA, Carlton JT, Palumbi SR (1995) Whaling effects on deep-sea biodiversity. Conserv Biol 9:462−464 Concluding remarks Cacchione DA, Rowe GT, Malahoff A (1978) Submersible investigation of outer Hudson submarine canyon. In: Mass depositions of salp carcasses represent a sig- Stanley DJ, Kelling G (eds) Sedimentation in submarine nificant pathway for the export of organic production canyons, fans and trenches. Dowden Hutchinson & Ross, from surface waters to the deep sea. Salp biomass in Stroudsburg, PA, p 42−50 −3 Childress JJ, Nygaard MH (1973) The chemical composition the Tasman Sea regularly exceeds 100 t km WW, of midwater fishes as a function of depth of occurrence with deposition events likely to export at least 16 t off southern California. Deep-Sea Res 20:1093−1109 km−2 of carbon, or 616 GJ km−2 of energy, to the ben- Clark MR (1985) The food and feeding of seven fish species thos every year. With higher organic content than from the Campbell Plateau, New Zealand. NZ J Mar Freshw Res 19:339−363 depositions of other gelatinous organisms, the input Clark MR, King KJ, McMillan PJ (1989) The food and feed- of large salp carcasses (salp-fall) is likely to be im - ing relationships of black oreo, Allocyttus niger, smooth portant to the nutritional ecology of the deep-sea oreo, Pseudocyttus maculatus, and eight other fish spe- benthos. cies from the continental slope of the south-west Chatham Rise, New Zealand. J Fish Biol 35:465−484 Clarke A, Holmes LJ, Gore DJ (1992) Proximate and ele- Acknowledgements. This research could not have been mental composition of gelatinous zooplankton from the completed without the kind and gracious assistance of S. Southern Ocean. J Exp Mar Biol Ecol 155:55−68 Mills and Dr. K. Schnabel from the Marine Invertebrate Col- Davenport J, Balazs GH (1991) Fiery bodies—are pyroso- lection at NIWA; M. Lewis from CSIRO; and Dr. M. Cryer mas an important component of the diet of leatherback and C. Loveridge from the Ministry of Fisheries, New turtles? Brit Herpetol Soc Bull 37:33−38 Zealand. We also thank M. Gall (UWS) for her assistance Domanski PA (1984) Giant larvae: prolonged planktonic lar- with the biochemical analysis of the specimens and A. Hart val phase in the asteroid Luidia sarsi. Mar Biol 80: (NIWA) for fish identification. We are grateful to the anony- 189−195 mous reviewers for their helpful comments that improved on Donnelly J, Torres JJ, Hopkins TL, Lancraft TM (1994) the original version of this manuscript. The New Zealand Chemical composition of Antarctic zooplankton during samples were collected under the Ocean Survey 20/20 austral fall and winter. Polar Biol 14:171−183 Chatham/Challenger Biodiversity and Seabed Habitat pro- Dubischar CD, Pakhomov E, Bathmann UV (2006) The tuni- ject, funded jointly by the New Zealand Ministry of Fish- cate Salpa thompsoni ecology in the Southern Ocean II. 174 Mar Ecol Prog Ser 491: 165–175, 2013

Proximate and elemental composition. Mar Biol 149: others (2013) Jelly biomass sinking speed reveals a fast 625−632 carbon export mechanism. Limnol Oceanogr 58:1113−1122 Dubois M, Gilles KA, Hamilton JK, Rebers PA, Smith F Lyle JM, Smith DC (1997) Abundance and biology of warty (1956) Colorimetric method for determination of sugars oreo (Allocyttus verrucosus) and spikey oreo (Neocyttus and related substances. Anal Chem 28:350−356 rhomboidalis) (Oreosomatidae) off south-eastern Aus- Duggins DO (1981) Sea urchins and kelp: the effects of short tralia. Mar Freshw Res 48:91−102 term changes in urchin diet. Limnol Oceanogr 26: Madin LP, Cetta CM, McAlister VL (1981) Elemental and 391−394 biochemical composition of salps (Tunicata: Thaliacea). Dunn MR, Griggs L, Forman J, Horn P (2010) Feeding habits Mar Biol 63:217−226 and niche separation among the deep-sea chimaeroid Madin LP, Kremer P, Wiebe PH, Purcell JE, Horgan EH, fishes Harriotta raleighana, Hydrolagus bemisi and Hydro- Nemazie DA (2006) Periodic swarms of the salp Salpa lagus novaezealandiae. Mar Ecol Prog Ser 407:209−225 aspera in the slope water off the NE United States: bio- Folch J, Lees M, Sloane Stanley GH (1957) A simple method volume, vertical migration, grazing, and vertical flux. for the isolation and purification of total lipides from ani- Deep-Sea Res I 53:804−819 mal tissues. J Biol Chem 226:497−509 Mitchell JK, Holdgate GR, Wallace MW, Gallagher SJ (2007) Gili JM, Rossi S, Pages F, Orejas C, Teixido N, Lopez- Marine geology of the Quaternary Bass Canyon system, Gonzalez PJ, Arntz WE (2006) A new trophic link be- southeast Australia: a cool-water carbonate system. Mar tween the pelagic and benthic systems on the Antarctic Geol 237:71−96 shelf. Mar Ecol Prog Ser 322:43−49 Moline MA, Claustre H, Frazer TK, Schofield O, Vernet M Godeaux J (1965) Observations sur la tunique des tuniciers (2004) Alteration of the food web along the Antarctic pelagiques. Rapp Comm Int Mer Médit 18:457–460 Peninsula in response to a regional warming trend. Glob Gooday AJ (2002) Biological responses to seasonally varying Change Biol 10:1973−1980 fluxes of organic matter to the ocean floor: a review. Moseley HN (1880) Deep-sea dredging and life in the deep JOceanogr58:305−332 sea. Nature 21:591−593 Grassle JF, Morse-Porteous LS (1987) Macrofaunal coloniza- Nakamura EL, Yount JL (1958) An unusually large salp. Pac tion of disturbed deep-sea environments and the struc- Sci 12:181 ture of deep-sea benthic communities. Deep-Sea Res Nishikawa J, Naganobu M, Ichii T, Ishii H, Terazaki M, Part A 34:1911−1950 Kawaguchi K (1995) Distribution of salps near the south Hill P (2009) Designing a deep-towed camera vehicle using Shetland Islands during austral summer, 1990−1991 with single conductor cable. Sea Technol 50:49−51 special reference to krill distribution. Polar Biol 15:31−39 Hirose E, Kimura S, Itoh T, Nishikawa J (1999) Tunic mor- Nodder SD, Pilditch CA, Probert PK, Hall JA (2003) Variabil- phology and cellulosic components of pyrosomas, dolio- ity in benthic biomass and activity beneath the Subtropi- lids, and salps (Thaliacea, Urochordata). Biol Bull 196: cal Front, Chatham Rise, SW Pacific Ocean. Deep-Sea 113−120 Res I 50:959−985 Hoeksema BW, Waheed Z (2012) It pays to have a big O’Driscoll RL (2004) Estimating uncertainty associated with mouth: mushroom corals ingesting salps at northwest acoustic surveys of spawning hoki (Macruronus novaeze- Borneo. Mar Biodiversity 42:297−302 landiae) in Cook Strait, New Zealand. ICES J Mar Sci Huntley ME, Sykes PF, Marin V (1989) Biometry and tropho- 61:84−97 dynamics of Salpa thompsoni Foxton (Tunicata, Thali- Parr Instrument Company (2008) 6200 Calorimeter Operat- acea) near the Antarctic Peninsula in austral summer, ing Instruction Manual No. 442M. Moline, IL 1983–1984. Polar Biol 10:59−70 Percy JA, Fife FJ (1981) The biochemical composition and Iguchi N, Kidokoro H (2006) Horizontal distribution of energy content of Arctic marine macrozooplankton. Arc- Thetys vagina Tilesius (Tunicata, Thaliacea) in the Japan tic 34:307−313 Sea during spring 2004. J Plankton Res 28:537−541 Perissinotto R, Pakhomov EA (1998) Contribution of salps to Ikeda T, Yamaguchi A, Matsuishi T (2006) Chemical compo- carbon flux of marginal ice zone of the Lazarev Sea, sition and energy content of deep-sea calanoid copepods Southern Ocean. Mar Biol 131:25−32 in the western North Pacific Ocean. Deep-Sea Res I Platt T, Irwin B (1973) Caloric content of phytoplankton. 53:1791−1809 Limnol Oceanogr 18:306−310 Kawahata H, Ohta H (2000) Sinking and suspended pa - Roe HSJ, Billett DSM, Lampitt RS (1990) Benthic/midwater rticles in the south-west Pacific. Mar Freshw Res 51: interactions on the Madeira Abyssal Plain; evidence 113−126 for biological transport pathways. Prog Oceanogr 24: Larson RJ, Harbison GR (1989) Source and fate of lipids in 127−140 polar gelatinous zooplankton. Arctic 42:339−346 Rowe GT, Staresinic N (1979) Sources of organic matter to Lebrato M, Jones DOB (2009) Mass deposition event of the deep-sea benthos. Ambio Spec Rep 6:19−23 Pyrosoma atlanticum carcasses off Ivory Coast (West Sherlock M, Kloser R, Williams A, Ryan T (2010) A combined Africa). Limnol Oceanogr 54:1197−1209 benthic grab and observation system for rapid assess- Lebrato M, Pahlow M, Oschlies A, Pitt KA, Jones DOB, ment of water column and seabed. In: Proc OCEANS’10 Molinero JC, Condon RH (2011) Depth attenuation of Institute of Electrical and Electronics Engineers Conf, organic matter export associated with jelly falls. Limnol Sydney, Australia, May 24−27, 2010, p 1−5 Oceanogr 56:1917−1928 Smith CR, Baco AR (2003) Ecology of whale falls at the Lebrato M, Pitt K, Sweetman A, Jones D and others (2012) deep-sea floor, Vol 41. Taylor & Francis, London Jelly-falls historic and recent observations: a review to Smith KL, Kaufmann RS (1999) Long-term discrepancy drive future research directions. Hydrobiologia 690: between food supply and demand in the deep eastern 227−245 North Pacific. Sci 284:1174−1177 Lebrato M, de Jesus Mendes P, Steinberg DK, Cartes JE and Smith CR, De Leo FC, Bernardino AF, Sweetman AK, Arbizu Henschke et al.: Salp carcasses as food-fall 175

PM (2008) Abyssal food limitation, ecosystem structure and climate change. Trends Ecol Evol 23:518−528 a Stockton WL, DeLaca TE (1982) Food falls in the deep sea: occurrence, quality, and significance. Deep-Sea Res I 29:157−169 Sweetman AK, Chapman A (2011) First observations of jelly-falls at the seafloor in a deep-sea fjord. Deep-Sea Res I 58:1206−1211 lenger Plateau, 2007, Plateau, lenger

Sweetman AK, Witte U (2008) Response of an abyssal 1631 2685 macrofaunal community to a phytodetrital pulse. Mar − Ecol Prog Ser 355:73−84 Thompson H (1948) Pelagic tunicates of Australia. Common- wealth Council for Scientific and Industrial Research,

Melbourne, p 136−139 es denote % contribution (WW) to Tranter DJ (1962) Zooplankton abundance in Australasian waters. Mar Freshw Res 13:106−142 Wacasey JW, Atkinson EG (1987) Energy values of marine benthic invertebrates from the Canadian Arctic. Mar Ecol Prog Ser 39:243−250 Wiebe PH, Madin LP, Haury LR, Harbison GR, Philbin LM (1979) Diel vertical migration by Salpa aspera and its 441 800 1600

potential for large-scale particulate organic matter trans- − port to the deep-sea. Mar Biol 53:249−255 Witte U, Pfannkuche O (2000) High rates of benthic carbon remineralisation in the abyssal Arabian Sea. Deep-Sea Res II 47:2785−2804 Wood RA (1991) Structure and seismic stratigraphy of the

western Challenger Plateau. NZ J Geol Geophys 34:1−9 1800 421 − Yamamoto J, Hirose M, Ohtani T, Sugimoto K and others (2008) Transportation of organic matter to the sea floor by carrion falls of the giant jellyfish Nemopilema nomurai in the Sea of Japan. Mar Biol 153:311−317 Young JW, Bradford RW, Lamb TD, Lyne VD (1996) Biomass

of zooplankton and micronekton in the southern bluefin 1000 1001 tuna fishing grounds off eastern Tasmania, Australia. − Mar Ecol Prog Ser 138:1−14 indicate contribution of salps to the haul. n = number of hauls of number = n haul. the to salps of contribution indicate 800 801 − values values Bold Challenger Plateau Bass Canyon 600 601 − haul excluding fish. excluding haul 500 501 − 3.38 92.35 9.54 12.00 25.03 13.22 2.64 1.90 5.27 12.49 (14.54) 22.23 (26.95) 23.74 (25.70) 17.43 (18.49) 18.87 (20.13) 20.6 (51.89) 30.3 (36.51) 63.1 (63.10) 55 (56.24) Pyrosoma atlanticum Percentage (wet weight, WW) of organisms and total haul weight (kg WW) across different depths from hauls performed on the Chal the performedon hauls fromdifferentacross depths WW) (kg weight haul total and organisms of WW) weight, (wet Percentage Sample includes n117104121111 Water depth (m)FishTunicates – salps 200 Tunicates–otherSpongesCnidariansEchinodermsPolychaetesSipunculids 0 (0)Molluscs 14.09 22.71 (26.44)Crustaceans 0.01 (0.01) 7.71 (8.98)Other 12.89 (15.62) 16.29 (18.96) 11.73 (14.22)Total haul 1.88 (2.28) 12.52 (13.55) (kg) 0 (0) 15.14 (16.39)a 1.34 (1.62) 14.94 (15.85) 17.5 0 (0) 12.01 (13.98) 8.46 (9.16) 0.16 (0.17) 11.85 (12.83) 21.67 (23.12) 6.53 (7.60) 6.58 (7.98) 28.42 (30.15) 15.2 (16.12) 4.48 (4.78) 2.6 (6.55) 8.87 (10.75) 0 (0) 4.06 (4.33) 10.77 (11.66) 13.57 (14.48) 8.15 (9.49) 7.61 0 (0) 10.2 (25.69) 4.81 (5.21) 23.8 (28.67) 9.98 (10.59) 0.1 (0.25) 16.98 (20.58) 0.5 (1.26) 29.9 (29.90) 0.02 7.75 (8.27) 7.47 (7.92)(0.02) 1 (1.20) 5.09 (5.51) 15.7 (16.05) 20.6 (24.82) 14.24 (15.19) 2.8 (3.37) 0 (0) 5.72 0.15 3.5 (8.82) (0.16) 0.65 (0.69) 0.9 (0.90) 1.1 (2.77) 0 (0) 0.5 (0.50) 14.6 (14.93) 8.93 (9.53) 2.8 (3.37) 1.1 (2.77) 1.2 (1.45) 0 (0) 6.28 0.4 (0.41) 0.9 (0.90) 0 (0) 0 (0) 0.5 (0.60) 3.2 (3.20) 0.1 (0.10) 0 (0) 0.6 (0.61) 57 0.6 (0.6) 0 (0) 11.4 (11.66) 0 (0) 17 0 (0) 0 (0) 0 (0) 0 0.7 (0.70) 0 (0) 2.1 Appendix 1. Appendix and in Bass Canyon, 2009. outside Values parentheses denote % contribution (WW) to haul including fish; values inside parenthes

Editorial responsibility: Marsh Youngbluth, Submitted: November 13, 2012; Accepted: June 17, 2013 Fort Pierce, Florida, USA Proofs received from author(s): September 13, 2013